针对基本的快速拓展随机树算法(rapidly-exploring random tree,RRT*)存在搜索随机性大、效率低、路径非最优的缺点,提出一种引入人工势场法算法(artificial potential field method,APF)和Douglas-Peucker算法的改进RRT*-APF-DP路径规划算法.在RRT*算法的采样点生成阶段引入变采样范围偏置搜索与步长自适应调整策略,融合重新设计的APF算法的引力与斥力函数,增强路径扩展导向性与绕过障碍物能力.采用重采样策略改进DP算法,优化避障代价与控制点数量.实验结果表明,本算法规划的避障路径满足机械臂的运动要求,且算法规划的避障路径代价、规划时间和路径控制节点数均得到有效改善.
Path planning for robot with improved RRT*-APF-DP algorithm
In response to the inherent shortcomings of the basic rapidly-exploring random tree(RRT*)algorithm,characterized by substantial search randomness,low efficiency,and non-optimal path generation,an improved path planning algorithm,denoted as RRT*-APF-DP,is proposed herein.This enhancement incorporates the artificial potential field(APF)method and the Douglas-Peucker al-gorithm.During the sample point generation phase of the RRT*algorithm,a variable sampling range biased search and adaptive step size adjustment strategy are introduced.This amalgamation is comple-mented by the integration of a redesigned APF algorithm featuring attractive and repulsive functions.The synergy of these components enhances the directed expansion of paths and the algorithm's capabil-ity to circumvent obstacles.Furthermore,the DP algorithm is refined through a resampling strategy,optimizing obstacle avoidance costs and the number of control points.Experimental results demonstrate that the obstacle-avoiding paths generated by the improved algorithm meet the motion requirements of the robotic arm.The algorithm effectively improves obstacle-avoidance path costs,planning time,and the number of path control nodes.
path planningrobotic armimproved-RRT*algorithmpath optimizationimproved artificial potential field methodDouglas-Peucker algorithms
万方数据
